This project is concerned with the development and/or evaluation of methodology for clinical psychopharmacology research and with statistical support of clinical investigations. The methodologic pursuits focus on data analysis and improved criteria for diagnosis and classification of patients in clinical psychopharmacology research. Work currently underway concerns evaluation of power and bias in several tests of significance for differences in response probabilities in 2 x 2 contingency tables. Two new statistics have been developed that appear to have less bias and greater power than previously available tests. Exact power tables are being constructed for all tests. Objective criteria for establishing a diagnosis of schizophrenia or depressive disorder have been proposed by several different authors. Work will be undertaken to compare the reliability and validity of the alternative diagnostic methods. Phenomenological (descriptive) classification will be compared with diagnostic classification regarding utility in defining patient populations in clinical psychopharmacology research. The computer laboratory will provide statistical support and data analysis for clinical investigators.